Curating Dataset Pipelines to Train Medical Chatbots on Early Sepsis Detection
International Journal of Artificial Intelligence and Green Manufacturing · 30.12.2025 · 1 (4)
DOI: 10.67119/0643wqzl ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Artificial Intelligence and Green Manufacturing, online erschienen am 30.12.2025. Verfasst von Arup Das und 6 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Themen: Sepsis
Abstract (Originalsprache)
Sepsis is a critical medical condition that arises when the body’s response to infection causes life-threatening organ dysfunction. Despite increasing awareness and the use of protocol-driven management strategies, early diagnosis remains a persistent challenge in clinical practice, especially in high- pressure settings such as emergency departments and ICUs. Nurses, as first responders, are crucial in identifying early signs, but often work under cognitive overload and ambiguity of the protocol. Large language models (LLMs) represent frontier neural network techniques that use self-supervised learning algorithms to process and understand human languages or text. This work focuses on building a robust data gathering pipeline in order to ultimately create an interactive clinical chatbot fine- tuned on sepsis-specific knowledge. The pipeline consists of a three-step process, namely lexical analysis, semantic analysis, and Q&A quality evaluation, that utilizes artificial intelligence to collect training data in novel ways. It provides a feasible and cutting-edge framework for LLM-based chatbot design and development.
Zitieren
APA
Das, A., Abecasis, M., Farrell, T., Sasson, I., Velandia, S. R., Tortorelli, B., & Wang, J. (2025). Curating Dataset Pipelines to Train Medical Chatbots on Early Sepsis Detection. International Journal of Artificial Intelligence and Green Manufacturing, 1(4). https://doi.org/10.67119/0643wqzl
Vancouver
Das A, Abecasis M, Farrell T, Sasson I, Velandia SR, Tortorelli B, et al. Curating Dataset Pipelines to Train Medical Chatbots on Early Sepsis Detection. International Journal of Artificial Intelligence and Green Manufacturing. 2025;1(4). doi:10.67119/0643wqzl
BibTeX
@article{Das2025,
author = {Das, Arup and Abecasis, Miriam and Farrell, Thomas and Sasson, Isaac and Velandia, Sophia Ramirez and Tortorelli, Brooke and Wang, Jiacun},
title = {Curating Dataset Pipelines to Train Medical Chatbots on Early Sepsis Detection},
journal = {International Journal of Artificial Intelligence and Green Manufacturing},
year = {2025},
volume = {1},
number = {4},
doi = {10.67119/0643wqzl}
}
EUVIMED